Lending Perspectives: It’s Time to Revisit Your Approach to Pattern Matching

credit score laptop gauge
By George Hofheimer

5 minutes

Credit unions can make banking more human by carving out experiments like this one.

This article is excerpted with permission from Hofheimer’s 2023 book, Banking on a Human Scale.

Technology is both a curse and a dream. The dream part is fairly obvious to most: You can take a handheld, wireless device and with the touch of a single button on a screen have a dozen buffalo chicken wings arrive at your door within 30 minutes. Less transformational examples of technology’s promise include predictive analytics, massive data storage and the remote delivery of financial services. 

The curse part may be less obvious, or as Scientific America states in this article, “We may be making ourselves dumber when we outsource thinking and rely on supposedly smart tech to micromanage our daily lives for the sake of cheap convenience.” 

Examples abound: Can you drive to a new location without the help of GPS? Can you remember who played the lead character in National Treasure without asking Alexa? When was the last time you went to the restroom without scrolling through Reddit or Fruit Ninja? In short, we have never in the course of human history been blessed with more access to more information, but I sincerely worry that we are losing knowledge (and becoming dumber) along the way. 

For credit unions, this shift in the use of technology is wrapped in the promise of more efficiency. However, this claim may be a bit spurious as U.S. credit unions’ operating expense is roughly the same as it was in 1991 with only a 10-basis point deviation over the past 30 years. Technology will play an oversized role in the future of consumer finance, but too much of a good thing may not be the best way forward. Credit unions have the opportunity to carve out non-technological experiments to make banking more human. Which brings me to the topic of pattern matching.

The Three Cs of Lending

The legacy of most credit unions (and banks for that matter) was lending according to the 3 Cs: capital, capacity and character. The capital and capacity variables have been largely outsourced to the credit bureaus. We can debate around the edges of their usefulness. On average, the systemization of these data holds value for both consumers and lenders. 

Character, on the other hand, has gone by the wayside except for commercial loans where part of the lending process involves inspection and evaluation of the management team’s abilities. Credit unions overwhelmingly lend to consumers and, with a few rare exceptions, totally automate the lending process without considering character, or more specifically, circumstances.

The following is a fictionalized (but approximately correct) example of a friend’s experience with a credit union. This individual had a fairly large ding on his credit report because of $15,000 in medical debt he incurred as an uninsured student at the local technical college three years earlier. At the time of his loan application, he was an emergency medical technician making $42,000/year with $2,500 in credit card debt, $900 in savings and a direct deposit with his credit union. His credit score (capital and capability) was 620. He was 28 years old and single. He wanted to buy a used car for $15,000. You have no information about his character. Based on your current lending criteria would you make a loan to this dude? If so, at what cost? 

I’ll cut to the chase and tell you this friend was turned down for the loan and referred by the credit union to a series of resources on improving one’s credit score in the form of outdated PDF brochures. He got the same pamphlets in the mail and a follow-up call from a member service representative asking how his experience was. Since he’s a passive-aggressive Midwesterner, he gave the credit union an “8” on their Net Promoter Score. He eventually obtained the loan from the used car dealership’s F&I department at approximately 15% APR, or a little over $500/month for a 2011 Subaru with 85,000 miles.

Remember the Fourth “C,” Character?

Now, what you don’t know about this person is that his character is unimpeachable. His medical debt was incurred intervening in a foiled assault on a mutual friend. While he is an EMT today, he was just accepted to medical school in New York, and he makes some cash playing guitar on the weekends at the local pub.  

But how would you know those things? Your systems are engineered to find a credit score, and that’s pretty much it. Credit unions have moved away from their legacy of character lending for a variety of really good reasons, but I can’t help but think your credit union would benefit if it rediscovered this differentiated approach. With a human department, perhaps?

Pattern matching is what you are currently doing with credit scoring. You create a hypothesis of patterns in data, collect observed data and determine if your hypothesis is correct with said data. But what if you applied pattern matching to character? Since character is a lot more squishy than debt and income measures, you will have to tap into the previous discussions on research methods and approaches. So, what might this look like in practice? Allow me to set up an experiment that you could run at your credit union. 

A Test of Character

Let’s start with the hypothesis that the credit union is missing lending opportunities because they don’t consider character in the credit decision-making process. To test this, randomly choose 25 members that apply for an automobile loan during a specific time frame at the credit union. Supplement their loan application with a simple request, “Please share any information about yourself that you think is important in helping XYZ Credit Union make this loan.” I’d provide examples of what applicants might wish to share, including extenuating circumstances, unique qualities/experiences and future plans.

Then, I would gather a diverse group of five credit union employees representing lending, member service, finance, marketing and branch operations to review these 25 applications. 

Finally, I’d track the loan performance of the people whose “character” put them into a loan approval when the traditional lending criteria did not. Repeat this process several times and you will likely identify a variety of patterns unique to the members you serve, thereby resulting in a potential strategic advantage over other lenders in your marketplace. Of course, at the end of this experiment, you’d want to test your hypothesis and report back on the findings.

George Hofheimer advises highly ambitious credit unions that want to change the world and the author of Banking on a Human Scale. The founder and principal of Hofheimer Strategy Advisors, he also is a co-founder and principal of The Strategy Circle.

Compass Subscription